Middle Eocene Trees of the Clarno Petrified Forest, John Day Fossil Beds National Monument, Oregon
Total Page:16
File Type:pdf, Size:1020Kb
PaleoBios 30(3):105–114, April 28, 2014 © 2014 University of California Museum of Paleontology Middle Eocene trees of the Clarno Petrified Forest, John Day Fossil Beds National Monument, Oregon ELISABETH A. WHEELER1 and STEVEN R. MANCHESTER2 1Department of Forest Biomaterials, North Carolina State University, Raleigh, NC 27605-8005 USA; elisabeth_ [email protected]. 2Florida Museum of Natural History, University of Florida, Gainesville, FL 32611 USA; steven@ flmnh.ufl.edu One of the iconic fossils of the John Day Fossil Beds National Monument, Oregon, USA, is the Hancock Tree—a permineralized standing tree stump about 0.5 m in diameter and 2.5 m in height, embedded in a lahar of the Clarno Formation of middle Eocene age. We examined the wood anatomy of this stump, together with other permineralized woods and leaf impressions from the same stratigraphic level, to gain an understanding of the vegetation intercepted by the lahar. Wood of the Hancock Tree is characterized by narrow and numerous vessels, exclusively scalariform perforation plates, exclusively uniseriate rays, and diffuse axial parenchyma. These features and the type of vessel-ray parenchyma indicate affinities with the Hamamelidaceae, with closest similarity to the Exbucklandoideae, which is today native to Southeast and East Asia. The Hancock Tree is but one of at least 48 trees entombed in the same mudflow; 14 others have anatomy similar to the Hancock Tree; 20 have anatomy similar to Platanoxylon haydenii (Platanaceae), two resemble Scottoxylon eocenicum (probably in order Urticales). The latter two wood types occur in the nearby Clarno Nut Beds. Two others are distinct types of dicots, one with features seen in the Juglandaceae, the other of unknown affinities, and the rest are very poorly preserved and of unknown affinity. Leaf impressions in and immediately below the layer containing the trees include the extinct genera Macginitiea and Platimeliphyllum (Platanaceae), and Trochodendroides (Saxifragales). Keywords: Eocene, Clarno Formation, paleobotany, fossil wood, Hamamelidaceae, Platanaceae, wood anatomy INTRODUCTION considered to have been deposited subsequent to the folding, The Nut Beds locality in the Clarno Formation of north- faulting, intrusion and erosion of the strata that include the central Oregon contains one of North America’s most diverse Nut Beds (his unit B). Whereas the Hancock Canyon lahar and well-documented assemblages of middle Eocene plants, contains large tuffaceous clasts up to 30 cm in diameter, with fruits, seeds, leaves, and woods (Scott 1954, Manchester clasts in the Nut Beds deposits generally do not exceed five 1981, Manchester 1994, Wheeler and Manchester 2002). cm. Although woods of the Nut Beds are highly abraded and About two km east of the Nut Beds in Hancock Canyon, evidently allochthonous (i.e., transported), those of Hancock Wheeler County, and in what is today the John Day Fossil Canyon are logs and stumps with minimal abrasion and may Beds National Monument, there is a petrified forest preserved represent more or less in situ forest intercepted by a lahar. in a volcanic lahar. Compared to the Nut Beds the quantity The botanical affinities of the Hancock Canyon trees have and diversity of the Hancock Canyon woods is low, but there not been examined in detail. Here we provide an overview of are logs and standing trunks up to 60 cm diameter contrast- the wood anatomical types occurring in the Hancock Canyon ing with the fragmentary eroded woods of the Nut Beds. area, and document associated leaf impressions. This informa- One of the large standing trunks, visible along the US Park tion will further our understanding of the diversity of woody Service’s Hancock Canyon trail, is named the Hancock Tree. plants in the Eocene of northwestern North America, as well This canyon, the tree, and the nearby field station operated as aid Monument personnel in reconstructing the paleoen- by the Oregon Museum of Science and Industry (OMSI), vironment of the John Day basin for their public exhibits. are named for Alonzo W. “Lon” Hancock, a devoted amateur paleontologist who collected extensively in John Day Basin MATERIALS AND METHODS in the 1950s and 60s, especially in the Clarno Formation Collections (Ashwill 1987). Samples were obtained during the summers of 1970 and Bestland et al. (1999) place the Hancock Canyon floras 1971, prior to the collecting restrictions imposed when this in the same subhorizon of the Clarno Formation as the Nut area became part of the John Day Fossil Beds National Monu- Beds and consider the two floras to be essentially coeval, ment. Specimens, slides, and detailed locality information are although the beds cannot be physically traced from one to archived at the Florida Museum of Natural History. Although the other due to an intervening dacite dome. On the other the woods occur in a single horizon, they are exposed in dif- hand, Hanson (1996, and pers. comm. 2014) maps the Han- ferent outcrops within a 0.2 km radius of the Hancock Tree cock Canyon flora in a higher stratigraphic level (his unit C), (N44.92884°, W120.415686°), with the following locality 106 PALEOBIOS, VOL. 30, NUMBER 3, APRIL 2014 numbers: UF48 (outcrop B); UF49 (outcrop C); UF68 and as listed in Gregory 1980, 1994) and the on-line Kew (outcrop D), UF69 (outcrop E). Samples are cataloged Micromorphology Database. with the 5-digit specimen number of the Florida Museum of Natural History at University of Florida (UF) preceded SYSTEMATIC PALEONTOLOGY by the UF locality number. Original field numbers (e.g., C1, The Hancock Tree indicating first sample collected from outcrop C) are also family: hamamelidaceae R. BR. 1818 indicated. Samples for thin sections were not taken from all suBfamily: exBucklandioideae Harms 1930 of the Hancock Canyon logs, and these logs are referred to cf. Hamamelidoxylon uniseriatum Wheeler and by their field number only. Manchester 2002 Sample preparation (Figs. 1–6) A diamond lapidary saw was used to cut thick sections Description—The growth rings are distinct and marked by (wafers) of transverse, tangential, and radial surfaces. One side radially flattened fibers, with vessels at the end of a growth of the wafer was smoothed to remove saw marks, and then ring being narrower than those in the earlywood of the suc- affixed to a glass side using 24-hour transparent epoxy. The ceeding growth ring (Figs. 1, 2). Vessels are predominantly sections were then ground to an initial thickness of about 50 solitary, vessel outline is oval to angular (Figs. 1, 2), with µm, either by hand on a lapidary grinding wheel, or with the a mean tangential diameter of 55 µm (sd =-11; the range is aid of a Buehler petrographic thin section grinding machine. 31–80 µm), there are 50–82 vessels per mm2; the vessel ele- Further fine grinding was done by hand, using a glass plate ments are 874–1273 µm long (average 1062, n=12). Perfora- and a slurry of 400 grade silicon carbide grit, until they were tion plates are exclusively scalariform, with 20–36 bars per thin enough (ca 30 µm) to allow seeing anatomical details perforation plate (Fig. 3), the intervessel pits are rare because with transmitted light microscopy. Cover slips were mounted the vessels are predominantly solitary, but some scalariform using Canada Balsam to improve clarity for light microscopy. pits have been observed. Vessel-ray parenchyma pits are elongate (scalariform) and apparently with reduced borders Descriptions (Fig. 5). Fibers are non-septate, with walls medium-thick to Our descriptions of the woods generally conform to the thick. Axial parenchyma is diffuse, with strands of 5–8 cells, IAWA Hardwood List (IAWA Committee 1989), with ves- mostly 8 cells per strand. Rays are exclusively uniseriate (Fig. sel diameter and ray height averages based on 25 measure- 4) and consist predominantly of procumbent cells 24–45 cells ments; rays per mm based on 10 counts; vessels per mm2 high (Fig. 6), with an average of 22 cells (sd =11). There from counts of vessels in 10 fields of view. Relationships to are 12–17 rays per mm. Crystals were not observed. Storied extant groups were investigated using the InsideWood web structure is absent. site (InsideWood 2004-onwards; Wheeler 2011) and refer- Specimen number: UF69-47252/E13, upright stump ences found on the Kew Micromorphology Bibliography approximately 50 cm diameter at base. web site. Diameters (diam) of the logs are presented based Comments—Five of the sectioned Hancock Canyon on measurements in the field. woods have anatomy similar to the Hancock Tree. They have similar qualitative features and overlap in quantitative vessel Determining affinities features (Table 1). Eight other logs were not sectioned, but The identification procedure was begun with searches appeared similar macroscopically, having many narrow ves- of the InsideWood web site (Wheeler 2011; InsideWood sels and narrow rays; these are UF49-47372/C2 (diam=34 2004-onwards) and included consultation of descriptions in cm), UF49-47374/C4 (diam =18 cm), UF49-47376/C6 “Anatomy of the Dicotyledons” (Metcalfe and Chalk 1950, (diam =13 cm), UF49-47379/C9 (diam =46 cm), UF49- Metcalfe 1987, Cutler and Gregory 1998). Subsequently, 473781/C11 (diam=7 cm), UF49-045385/C15 (diam =12 comparisons were made to extant wood samples in the Bailey- cm), UF48-47351/B1 (diam=11 cm), UF69-47253/E14 Wetmore Laboratory of Plant Anatomy and Morphology (diam =35 cm). Stem diameters for this wood type range of Harvard University, Universiteit Leiden Branch of the from 7 to 50 cm. Nationaal Herbarium Nederlands, University of Utrecht, Affinities—Members of the Hamamelidaceae (Chunia Jodrell Laboratory of the Royal Botanic Gardens, and the H.T. Chang and Exbucklandia R.W. Br., subfamily Ex- David A. Kribs (NC State University) wood collections and bucklandoideae; Fothergilla L. and Hamamelis L., subfamily illustrations and descriptions in standard references (Ilic 1991; Hamamelidoideae Reinsch; Pentaphylacaceae Engler (Cleyera / Figures 1–6.